A four-partner accountancy practice in the south-west carries £1m of Professional Indemnity cover. Half of its fee income comes from owner-managed-business clients; the other half is split between a small audit portfolio and corporate finance lead-advisory work on transactions in the £1m to £10m bracket. The practice has been at £1m of cover for years because that is what the regulator's sliding scale required, and the renewal has rolled over each year.
A client sells a business for £4m. Six months later the buyer discovers what they describe as a material misstatement of working capital in the management accounts the firm signed off. The claim notified is for £900,000 plus the buyer's legal costs.
The £1m policy responds, but by the time defence costs are spent the practice's headroom is uncomfortably thin and the partners are personally exposed to anything that runs over. The minimum was the wrong number. Not because the regulator got it wrong — the regulator sets a floor that the profession as a whole can absorb — but because the floor doesn't reflect the practice's individual exposure profile.
This article is about how to think about that number for your own practice. It does not give you a single right answer because there isn't one; it gives you the questions to work through so that the figure you settle on is defensible to your partners, your insurer, and your professional body.
Start with the regulator's floor, but only as the floor
ICAEW's Professional Indemnity Insurance Regulations (revised with effect from 1 September 2024) require a minimum limit of indemnity of £2m for any one claim and in the aggregate. For firms with gross fee income below £800,000, the minimum is the greater of 2.5 times gross fee income or £250,000. Firms above £50m of gross fee income are not required to hold qualifying insurance but must have appropriate arrangements ICAEW monitors. ACCA's framework uses different figures — broadly, the greater of 2.5 times total income or £100,000 below £600,000 of income, and at least £1.5m at or above £600,000 — so a dual-registered firm must meet whichever bar is higher.
The regulator's minimum is set to be enough to make most plausible claims survivable for most plausible practices. It is not an estimate of your real risk. Many of the practices that come unstuck in claims were carrying exactly the regulator's minimum and discovered, after the fact, that it did not reflect the size of the engagements they were actually doing.
The three questions that actually size the limit
The right limit for your practice depends on three things, in order of importance.
The first is the maximum financial exposure on your most exposed live engagement. Take the three largest engagements on your books right now. For each, ask: if the work I am signing off is materially wrong, what is the worst-case financial loss the client or a third party would suffer? For a set of audited accounts, that's the difference between the reported numbers and reality. For a transaction-support engagement, it's the difference between what the buyer paid and what the business is actually worth. For a tax-planning structure, it's the tax-plus-interest-plus-penalties exposure if HMRC successfully challenges. Your PI limit should comfortably exceed the largest of those three numbers, with headroom for defence costs.
The second is the typical sector and structural profile of your client base. A practice acting predominantly for OMB clients in the £500k-to-£5m turnover bracket carries lower claim-severity risk than a practice acting for listed plcs, financial services firms, regulated pension schemes, or public-sector bodies. The same engagement is a bigger PI exposure if the client is sophisticated, well-advised, and has the legal budget to sue properly.
The third is the structural shape of the policy. A £2m "any one claim" limit with an unlimited aggregate covers very differently from a £2m "in the aggregate" limit where one settled claim exhausts your cover for the whole year. Most accountancy PI policies will be written as any-one-claim with some aggregate cap; the relationship between the per-claim figure and the aggregate figure is part of what you are buying.
Working backwards from worst-case scenarios
For an owner-managed-business-focused practice doing accounts prep, personal and corporate tax, and limited transaction work, a reasonable mental model is: my largest single engagement is probably the accounts and tax for my biggest corporate client; the worst-case exposure on that engagement is something like one year's audit/accounts-related claim value. For most such practices that lands in the £250,000-to-£750,000 range, and a £1m limit covers that with sensible headroom.
For a practice with significant tax-advisory work — EIS, R&D, capital allowances, restructurings, residency advice — the exposure is uncapped in principle but commonly sits in the £500,000-to-£2m range per engagement once HMRC interest and penalties are included. Practices in this profile typically buy between £2m and £5m.
For a practice doing audit work, particularly in regulated or listed-company territory, the exposure can be much higher because the loss flowing from a misstated set of audited accounts can be very large. Audit firms acting for any plc or any financial-services-regulated entity are typically buying in the £5m-and-up range, often with separately-underwritten higher layers above their primary cover.
These are not rules. They are rules-of-thumb that brokers and underwriters use as a starting point for the conversation about what your actual exposure looks like.
Excess — pay your way down, sensibly
Excess is the amount you bear before the policy responds. Under ICAEW's regulations the maximum permitted aggregate excess is the higher of £3,000 or 3% of the firm's gross fee income. Within that constraint, increasing your excess typically reduces your premium, sometimes meaningfully, because you are taking the high-frequency low-severity claims off the insurer's book.
The trade-off is straightforward: every claim that lands on your file you fund yourself up to the excess. A practice with a clean five-year history and a robust claims-handling discipline can usually absorb a higher excess profitably. A practice that has had two or three modest claims in recent years probably should not. The decision is partner-by-partner and year-by-year.
Aggregate caps — the question most practices don't ask
Aggregate is the total amount the policy will pay across all claims in the policy year. A £2m "any one claim, £2m aggregate" policy and a £2m "any one claim, £4m aggregate" policy carry the same per-claim cover but quite different total annual cover.
For practices with low claim frequency this rarely matters — you might go a decade without a notified claim. For practices with higher-volume client books, particularly where one underlying error could affect many clients at once (a software-driven payroll error, a tax-advice approach used across multiple clients, a common audit methodology issue), the aggregate matters a great deal. If one underlying root cause produces six related claims, an aggregate cap can be reached fast.
We routinely recommend practices look at the aggregate-to-per-claim ratio when comparing quotes. A higher aggregate is often inexpensive to add and disproportionately valuable when it matters.
Reinstatement provisions
Some policies offer a reinstatement of the limit after a claim has eroded it. Where available, this can be valuable for practices with one large claim that exhausts the per-claim limit but leaves the practice trading and exposed to further notifications during the same policy year. Reinstatement is more common at higher policy bands and usually comes with conditions; it is worth asking about, particularly if you carry a single-layer policy at a relatively low limit.
The role of higher layers
For practices buying above £5m of cover, the policy is typically structured as a primary layer (often £2m or £5m) plus excess layers above. Different insurers may sit on different layers, and the terms on the excess layers may differ from the terms on the primary. The wording differences matter — an excess layer that doesn't follow the primary's wording may leave you exposed at the interface between layers.
Mid-market practices rarely need to worry about layered structures, but anyone considering a step up from £1m to £5m or beyond should make sure the broker explains the layer structure and whether the layers are written follow-form (matching the primary) or on standalone wordings.
How your premium relates to your limit — not as linearly as you might think
Doubling your limit from £1m to £2m does not double your premium. PI premium curves are non-linear: the marginal cost of buying additional cover above the regulator's minimum is usually modest compared with the cost of the primary layer itself, because the claims that would reach into the higher layers are statistically much rarer than the claims that sit within the primary.
For a typical OMB-focused practice, moving from £1m to £2m might add 15-25% to the premium. Moving from £2m to £5m a similar increment. The ratio shifts when you cross into specifically high-risk engagement types (large audits, complex tax planning, listed-company transaction work), where the upper layers become substantially more expensive to fund.
This is one of the reasons many practices end up under-covered: the marginal cost of one more million is small, but the marginal cost of being half a million short on a claim is enormous.
A simple decision framework
If you want a decision framework rather than a number:
Look at your largest live engagement and ask the worst-case exposure question. Make sure your per-claim limit comfortably exceeds it with at least 30-40% headroom for defence costs. Look at the engagement-mix shift since your last renewal and adjust upward if you have moved into bigger or more sophisticated client work. Look at your claims history; if you have had a notification, the renewal market will be tighter and the value of extra headroom is higher. Check the aggregate-to-per-claim ratio and ask whether one root-cause-driving multiple claims is plausible in your practice. Decide your excess based on your cash-flow tolerance and your claims history. Then talk to your broker.
What Apex does on a sizing conversation
We work through these questions with you ahead of renewal, present the resulting profile to several participating insurers, and bring back quotes that reflect the cover-shape you have decided on. We do not have a quota with any insurer and we do not push a particular limit because it pays us better; the broker remuneration is the same shape regardless of which insurer wins the renewal. What we add is the negotiation, the wording comparison, and the documentation — the parts of the renewal that are most likely to matter if a claim subsequently arrives.
Our broader explainer on accountants' PI cover is the Accountants PI Insurance UK Guide 2026. Once you are ready to talk about your renewal, the accountants sector page or contact page is the place to start.